Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{188201,
        author = {Sandip Basnet and Dr. Mandeep Kaur},
        title = {Machine Learning and Longitudinal Biomarkers: Thematic Insights into Predicting Alzheimer Disease Progression},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {7},
        pages = {2103-2109},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=188201},
        abstract = {Alzheimer's disease (AD) is the most prevalent cause of dementia globally, with more than 55 million individuals affected, and cases estimated to almost double by 2050. The heterogeneous course of AD poses major obstacles in early diagnosis, prognosis, and treatment planning. Whereas conventional diagnostic procedures depend on neuroimaging, cerebrospinal fluid biomarkers, and neuropsychological testing, these frequently cannot make predictions for early courses of disease. Machine learning (ML) advances have brought new possibilities for utilizing longitudinal biomarker data to model and predict AD progression with greater accuracy. This paper is a secondary data-based thematic analysis of the existing literature, datasets, and frameworks in AD prediction. Based on published research, especially using large-scale efforts like the Alzheimer's Disease Neuroimaging Initiative (ADNI), the article highlights five essential themes: (1) temporal dynamics of biomarker prediction, (2) missing and incomplete data management, (3) integration of multimodal biomarkers, (4) model interpretability and clinical acceptance, (5) generalizability and real-world application and (6) Clinical Integration and Deployment Challenges. Each theme highlights distinctive opportunities and challenges in the use of ML for predicting AD progression. The review points out how temporal models such as LSTM and Transformer networks have improved sequential biomarker interpretation and imputation and generative methods have reduced missing data problems. Integration of multimodal MRI, PET, CSF, and genetic markers has enhanced predictive accuracy, although transparency, overfitting, and clinical uptake issues continue to exist. The research concludes that explainable, multimodal, and federated ML models are the AD prognosis of the future, harmonizing computational accuracy with ethical and clinical considerations.},
        keywords = {Alzheimer disease (AD), Machine Learning, Alzheimer's Disease Neuroimaging Initiative},
        month = {December},
        }